A Bayesian network classification methodology for gene expression data.

We present new techniques for the application of a Bayesian network learning framework to the problem of classifying gene expression data. The focus on classification permits us to develop techniques that address in several ways the complexities of learning Bayesian nets. Our classification model reduces the Bayesian network learning problem to the problem of learning multiple subnetworks, each consisting of a class label node and its set of parent genes. We argue that this classification model is more appropriate for the gene expression domain than are other structurally similar Bayesian network classification models, such as Naive Bayes and Tree Augmented Naive Bayes (TAN), because our model is consistent with prior domain experience suggesting that a relatively small number of genes, taken in different combinations, is required to predict most clinical classes of interest. Within this framework, we consider two different approaches to identifying parent sets which are supported by the gene expression observations and any other currently available evidence. One approach employs a simple greedy algorithm to search the universe of all genes; the second approach develops and applies a gene selection algorithm whose results are incorporated as a prior to enable an exhaustive search for parent sets over a restricted universe of genes. Two other significant contributions are the construction of classifiers from multiple, competing Bayesian network hypotheses and algorithmic methods for normalizing and binning gene expression data in the absence of prior expert knowledge. Our classifiers are developed under a cross validation regimen and then validated on corresponding out-of-sample test sets. The classifiers attain a classification rate in excess of 90% on out-of-sample test sets for two publicly available datasets. We present an extensive compilation of results reported in the literature for other classification methods run against these same two datasets. Our results are comparable to, or better than, any we have found reported for these two sets, when a train-test protocol as stringent as ours is followed.     

Ouabain on active transepithelial sodium transport in frog skin: studies with microelectrodes

Studies were done with isolated frog skin to determine the effects of 10(-4) M ouabain on the electrophysiological parameters of outer and inner barriers of the Na-transporting cells. Microelectrodes were used to impale the skins from the outer surface to determine the intracellular voltages (Vsco) under conditions of short-circuiting and under conditions where a voltage clamp was used to vary the transepithelial voltage, VT. From this, the electrical resistances of outer (Rfo) and inner (RI) barriers were estimated. In addition, the driving force for active transepithelial Na transport (ENa = E'1) was estimated from the values of VT when the Vo = 0 mV (Helman and Fisher. 1977. J. Gen. Physiol. 69: 571-604). Studies were done with skins bathed with the usual 2.4 meq/liter [K]i in the inner solution as well as with reduced [K]i of 0.5 and 0 meq/liter. Characteristically, the responses to ouabain could be described by an initial rapid phase (5-10 min) during which time the Ri was increased markedly and the E'1 was decreased from control values. Thereafter, during the slow phases of the response, the resistances of both outer and inner barriers increased continuously and markedly with time leading ultimately to essentially complete inhibition of the short-circuit current. Similar studies were done with skins exposed to 10(-4) M amiloride in the outer solution. Although estimates of Ri could not be obtained under these conditions, the effects on the Vsco and E'1 were similar to those observed for the Na-transporting skins. However, the magnitudes of the effects were less and relatively slower than observed for the Na-transporting skins. The results of these studies were analyzed within the context of a proposed electrical model that takes into account the observation that the magnitude of the voltage at the inner barrier appears to exceed the equilibrium potential for K especially when transepithelial Na transport is inhibited at the apical barrier of the cells.     

CXCR4 promotes oral squamous cell carcinoma migration and invasion through inducing expression of MMP-9 and MMP-13 via the ERK signaling pathway

The increased migration and invasion of oral squamous cell carcinoma cells are key events in the development of metastasis to the lymph nodes and distant organs. Although the chemokine receptor CXCR4 and its ligand, stromal cell-derived factor-1α, have been found to play an important role in tumor invasion, its precise role and potential underlying mechanisms remain largely unknown. In this study, we showed that knockdown of CXCR4 significantly decreased Tca8113 cells migration and invasion, accompanied with the reduction of MMP-9 and MMP-13 expression. Inhibition of ligand binding to CXCR4 by a specific antagonist TN14003, also led to reduced cancer cell migration and invasion. Because the degradation of the extracellular matrix and the basement membrane by proteases, such as matrix metalloproteinases (MMP) is critical for migration and invasion of cancer cells, we investigated the expression of several MMPs and found that the expression of functional MMP-9 and MMP-13 was selectively decreased in CXCR4 knockdown cells. More importantly, decreased cell migration and invasion of CXCR4 knockdown cells were completely rescued by exogenous expression of MMP-9 or MMP-13, indicating that the two MMPs are downstream targets of CXCR4-mediated signaling. Furthermore, we found the level of phosphorylated extracellular signal-regulated kinase (ERK) was significantly decreased in CXCR4-silenced cells, suggesting that ERK may be a potential mediator of CXCR4-regulated MMP-9 and MMP-13 expression in Tca8113 cells. Taken together, our results strongly suggest the underlying mechanism of CXCR4 promoting Tca8113 migration and invasion by regulating MMP-9 and MMP-13 expression perhaps via activation of the ERK signaling pathway.     

Genes encoding A-type flavoproteins are essential for photoreduction of O2 in cyanobacteria

O(2) photoreduction by photosynthetic electron transfer, the Mehler reaction, was observed in all groups of oxygenic photosynthetic organisms, but the electron transport chain mediating this reaction remains unidentified. We provide the first evidence for the involvement of A-type flavoproteins that reduce O(2) directly to water in vitro. Synechocystis sp. strain PCC 6803 mutants defective in flv1 and flv3, encoding A-type flavoproteins, failed to exhibit O(2) photoreduction but performed normal photosynthesis and respiration. We show that the light-enhanced O(2) uptake was not due to respiration or photorespiration. After dark acclimation, photooxidation of P(700) was severely depressed in mutants Deltaflv1 and Deltaflv3 but recovered after light activation of CO(2) fixation, which gives P(700) an additional electron acceptor. Inhibition of CO(2) fixation prevented recovery but scarcely affected P(700) oxidation in the wild-type, where the Mehler reaction provides an alternative route for electrons. We conclude that the source of electrons for O(2) photoreduction is PSI and that the highly conserved A-type flavoproteins Flv1 and Flv3 are essential for this process in vivo. We propose that in cyanobacteria, contrary to eukaryotes, the Mehler reaction produces no reactive oxygen species and may be evolutionarily related to the response of anaerobic bacteria to O(2).     

Autoregulation of apical membrane Na+ permeability of tight epithelia. Noise analysis with amiloride and CGS 4270

Noise analysis of the Na+ channels of the apical membranes of frog skin bathed symmetrically in a Cl-HCO3 Ringer solution was done with amiloride and CGS 4270. Tissues were studied in their control states and after inhibition of transepithelial Na+ transport (Isc) by addition of quinine or quinidine to the apical solution. A critical examination of the amiloride-induced noise indicated that the single channel Na+ currents (iNa) were decreased by quinine and quinidine, probably because of depolarization of apical membrane voltage. Despite considerable statistical uncertainty in the methods of estimation of the Na+ channel density with amiloride-induced noise (NA, see text), the striking observation was a large increase of NA with amiloride inhibition of the rate of Na+ entry into the cells. NA was increased to 406% of control, whereas Isc was inhibited to 8.6% of control by 6 microM amiloride. Studies were done also with the Na+ channel blocker CGS 4270. Noise analysis with this compound was advantageous, permitting iCGSNa and NCGS to be measured in individual tissues with a relatively small inhibition of Isc. As with amiloride, inhibition of Isc with CGS 4270 caused large increases of the Na+ channel density (approximately 200% at approximately 35% inhibition of the Isc). Quinine and quinidine caused an approximately 50% increase of Na+ channel density while inhibiting iNa by approximately 60-70%. As inhibition of Na+ entry leads to an increase of Na+ channel density, a mechanism of autoregulation appears to be a major factor in adjusting the apical membrane Na+ permeability of the cells.     

Na+ and K+ transport at basolateral membranes of epithelial cells. III. Voltage independence of basolateral membrane Na+ efflux

Na+ efflux across basolateral membranes of isolated epithelia of frog skin was tested for voltage sensitivity. The intracellular Na+ transport pool was loaded with 24Na from the apical solution and the rate of isotope appearance in the basolateral solution (JNa23) was measured at timed intervals of 30 s. Basolateral membrane voltage was depolarized by either 50 mM K+, 5 mM Ba++, or 80 mM NH+4. Whereas within 30 s ouabain caused inhibition of JNa23, depolarization of Vb by 30-60 mV caused no significant change of JNa23. Thus, both pump-mediated and leak Na+ effluxes were voltage independent. Although the pumps are electrogenic, pump-mediated Na+ efflux is voltage independent, perhaps because of a nonlinear relationship between pump current and transmembrane voltage. Voltage independence of the leak Na+ efflux confirms a previous suggestion (Cox and Helman, 1983. American Journal of Physiology. 245:F312-F321) that basolateral membrane Na+ leak fluxes are electroneutral.     

Psyche,soma, and society: the social construction of psychosomatic disorders

This pilot-study examines the self-perceptions, and explanatory models, of 42 patients with either respiratory or gastrointestinal psychosomatic disorders. For several reasons, these disorders comprise an anomalous category within the biomedical model. It is suggested that clinicians explain their chronic, unpredictable course by psychologization - — shifting responsibility for etiology, exacerbations or therapeutic failure to patients emotions, personality, or lifestyle. Evidence is presented that psychologization is socially constructed, in clinical encounters over time. Patients respond to this process by reifying pathogenic emotions, personality traits, or malfunctioning body parts, and thus separating them from an idealized concept of the social self. It is also suggested that patients with gastrointestinal or respiratory conditions differ in their self-perceptions and explanatory models: a proportion of patients in each group organize their experiences around a central natural symbol — respiration or digestion/excretion. These 2 images link physiological experiences to concepts of pathogenic emotions or personality, physical weakness, and types of social relationships.